“China is going to win the AI race.” The remark sent ripples through Silicon Valley and beyond when Jensen Huang, CEO of US chipmaking giant Nvidia, made it at an AI summit in London earlier this month. Huang, whose company dominates the global AI chip market, later softened his position, saying China is merely “nanoseconds behind America.”

But Greg Slabaugh, Professor of Computer Vision and AI at Queen Mary University of London, is convinced that China has “already won” the AI race. And he made a startling revelation to back up his claim: of all the papers presented at the 2025 International Conference on Computer Vision in Hawaii, half were authored by Chinese researchers – far outstripping the US at 17%. Factor in Chinese nationals working abroad, and the gap would widen even further.

Three years before Huang’s candid acknowledgement, the Australian Strategic Policy Institute had reported that China leads in 57 out of 64 critical technologies – from quantum sensors and AI to robotics and semiconductors – while the US maintains an edge in far fewer areas, such as biotechnology and aerospace. This marks a dramatic reversal from 2003 to 2007, when the US led in 60 of 64 technologies and China in just three. Beijing’s current dominance stems from a high-impact research ecosystem in which, in some fields, it holds something close to a near-monopoly.

This meteoric rise, especially in AI, couldn’t be stymied by US efforts to limit China’s access to advanced chips and manufacturing equipment, which were intended to maintain America’s edge in the sector. Unlike past advantages based on cheap labour or scale, China’s AI lead is structural, built on concerted strategy, coordinated investment, and an energy ecosystem optimised for massive computational growth.

The AI revolution is, at its core, a revolution of power – in both senses of the word. Training the largest data models requires a huge computing capacity, which is powered by electricity on a colossal scale. By the decade’s end, experts say, AI data centres could consume more power than some mid-sized nations. And China holds a decisive edge in this high-voltage contest. Its subsidised electricity, flexible regulation, and capacity to execute large-scale projects at rapid speed have led to the mushrooming of AI infrastructure nationwide. From data-centre clusters in Inner Mongolia to renewable-powered server farms in Sichuan, Beijing has built an energy foundation capable of sustaining AI’s exponential growth. On the contrary, US tech giants are increasingly hamstrung by a growing web of constraints. The American electricity grid is old and fragmented, creating logistical and regulatory bottlenecks. Microsoft has conceded that energy shortages are slowing the expansion of its data centre. Chinese authorities, meanwhile, have turned energy planning into a national security priority – integrating AI, cloud computing, and grid modernisation into one strategic blueprint.

While Western firms like OpenAI and Anthropic pursue closed, commercial AI models, Chinese developers have doubled down on open-weight systems – models whose trained parameters are freely available. The result has been an open-source explosion that is transforming global software development. Chinese open-source AI downloads have now surpassed those from the US, according to venture capital firm a16z. Companies such as DeepSeek, MiniMax, Z.ai, and Moonshot are releasing high-performance models at a fraction of US prices.

China’s innovation often lies not in raw capability, but in accessibility and cost efficiency. Airbnb CEO Brian Chesky recently revealed that his company had replaced OpenAI’s ChatGPT with Alibaba’s Qwen model, calling it “fast and cheap.” Chamath Palihapitiya, CEO of Social Capital, said his firm has switched to Moonshot’s Kimi K2, describing it as “way more performant” than American rivals.

Conflicting approaches are at play here. The United States, by its own admission, wants to maintain global leadership in AI to secure its economic competitiveness. China, on the other hand, pushes for democratisation of AI, promoting open cooperation, capacity building for developing nations, and an “AI for the public good.” With this approach, Beijing has flipped one of Washington’s strategic levers. American export controls – meant to slow Chinese progress by denying access to cutting-edge chips – have instead spurred Chinese firms to build leaner models that run on older hardware. “They’ve actually encouraged Chinese companies to be more resourceful,” said AI researcher Toby Walsh. “It’s exactly what happened with solar panels – constraints made them smarter and cheaper.”

The story doesn’t end there. One of the most consequential areas of Chinese dominance is remote sensing: the science of gathering data from a distance through satellites, drones, and advanced sensors. A recent global analysis of 126,000 peer-reviewed papers found that China produced nearly 47% of all remote sensing research between 2021 and 2023. The American share, which stood at 88% during the Cold War, has fallen to just 9%. The patent landscape tells the same story. Among the top 19 global patent filers in remote sensing between 2021 and 2023, Chinese institutions accounted for 62%. Why does this matter? Because remote sensing underpins nearly every next-gen technology – from self-driving cars and smart cities to climate modeling and precision agriculture. Whoever controls the sensors, data flows, and analytic algorithms effectively controls the informational foundation of modern economies.

That said, China’s leap was no accident. Since the early 2000s, Beijing has strategically targeted the field for heavy investment under national programmes like the “973 Plan,” pairing state funding with private enterprise. The result: a vast ecosystem of universities, startups, and ministries working in concert. The US, by contrast, has relied heavily on NASA and the private sector. But fragmented research funding, bureaucratic inertia, and inconsistent industrial policy have eroded its early lead. When one country produces nearly half of global output in a strategic domain, and controls most patents and funding, it is shaping the next generation of value chains.

China’s stratospheric rise extends far beyond data and algorithms. In sector after sector, Western companies find themselves being out-produced, out-priced, and out-innovated. In automobiles, Chinese brands are redefining global competition. In 2024, Chinese carmakers captured 7.4% of all passenger car sales in Europe, nearly doubling their share within a year. EV maker Leapmotor posted a staggering 7,000% jump in sales, while BYD and Chery continue their European expansion with EVs far more affordable than Western models. This is why Ford CEO Jim Farley recently issued a blunt warning: “They have enough production capacity in China to serve the entire North American market.”

The same dynamic plays out in wind power, where Chinese manufacturers like Goldwind, Envision, and Mingyang now occupy the top four global slots – pushing Western rivals Siemens Energy, GE, and Vestas down the rankings. Chinese turbines are up to 50% cheaper, thanks to economies of scale and domestic demand that dwarfs anything in Europe or the US.

Having said that, all is not lost for the West, particularly the US, which still dominates the premium end of AI, biotechnology, and aerospace. Yet Washington must rethink its approach: instead of trying to slow China’s rise through export controls and strategic containment, it should focus on large-scale investment in energy infrastructure, R&D, and education. At the same time, it needs to face the new reality.

The writer is an independent journalist with special interest in geoeconomics